Trained to investigate early stopping for fine-tuning the language model. We have decided to early stop at 5 epochs for future fine-tuning efforts.